Currently available therapies for advanced melanoma are inadequate. No therapy has been shown to prolong survival. Since the identification of BRAF mutations in the majority "ofcases of melanoma, targeting the MAP kinase pathway has become central to novel therapeutic strategies in melanoma. Sorafenib, a novel serine threonine kinase inhibitor with potency against BRAF, is the only small molecule inhibitor of this target to have advanced beyond phase I trials. Based on the results of phase II trials of sorafenib in combination with chemotherapy in melanoma, a randomized, placebo-controlled phase III trial (E2603) is being conducted through ECOG to compare the efficacy of sorafenib, carboplatin and paclitaxel (experimental arm) to carboplatin and paclitaxel (control arm) among patients with unresectable stage III or IV melanoma. The submission of tumor blocks for participants in E2603 is required and will constitute an invaluable resource. In order to identify patient subsets that are most likely to respond and those for which other investigational therapies may be of higher priority, biomarker analyses in the context of large-scale clinical trials are essential. The complexity of genetic changes in melanoma is an area of active research and recent data suggest that genetic changes in three interacting pathways (RAS signaling [MAPK, PI3K/Akt], p16-CDK4-Rb and p53) play important roles in melanoma. The genetic changes in these pathways are not independent and generally display pathway 'exclusivity'. Thus, to understand the determinants of response for patients treated in E2603 we need to build a complete mutational profile, taking into account, different levels of a single pathway as well as multiple simultaneously affected pathways. We propose three approaches to identify genetic changes as they relate to response. Specific Aim 1 focuses on genetic changes in genes known to be important in melanomagenesis; molecular studies will be done in 550 participants divided between the two arms. Specific Aim 2 uses array based comparative genomic hybridization (aCGH) to fully characterize genomic changes as they are associated with response in E2603 in 200 melanomas (50 responders and 50 non-responders on each arm). Specific Aim 3 focuses on the identification of amplified genes in non-responders as potential targets for the next set of clinical trials using expression profiling and aCGH with validation in functional studies. These aims will provide valuable information about selecting patients with melanoma for targeted therapies and provide a basis for further therapeutic development.